The new user interface is in preview!

Want to check it out? Click here! (If you don't like it, you can still switch back)

League Forums

Main - General MFN Discussion

Re: QB Accuracy With Maths Because I like Maths

By setherick
10/21/2016 7:55 pm
If you have seen it, JDB provided the magic formula for passing accuracy in the main forums. Well, he admitted at all players start with 100% knowledge of the best place to throw the ball and then there are a series of penalties that are applied to that placement. Here's the formula and thread:

(80% accuracy attribute + 20% experience*) x (reduction based on scramble skill/movement) x (reduction based on duress) x (reduction based on distance) = accuracy for pass

* "experience" is actually a more calculated formula that takes into account the player's positional experience, his experience with both the offensive and defensive plays, his fatigue, and crowd noise. His positional experience (what you see on his player card) carries the greatest weight, followed by his fatigue, followed by his play experience for each play, followed by the crowd noise.

https://private75.myfootballnow.com/community/thread/1/2287?page=2#15114


I'm most interested in that first set of parameters. And specifically when we reach a point of negative returns for accuracy and how drastic those negative returns are.

That footnote is the most important thing for the set of calculations I'm going to layout. The way I read that footnote, a player will never hit 100 experience because they will never have 100 knowledge of their play or a play they are playing against. There is enough anecdotal evidence to show that knowledge is gained too slowly to get there. So for the next set of calculations, 80 will be used as the max experience a player can obtain.

For clarity sake, setting the ceiling for experience at 80 will mean that accuracy attribute will never reach 100. The max will be 96 (80% x 100 Accuracy + 20% x 80 Experience).

The point that we reach negative returns is 80 with an experience ceiling of 80.

Accuracy Experience Adj Accuracy Difference
100 80 96 -4
95 80 92 -3
90 80 88 -2
85 80 84 -1
80 80 80 0
75 80 76 1
70 80 72 2
65 80 68 3
60 80 64 4
55 80 60 5
50 80 56 6
45 80 52 7
40 80 48 8
35 80 44 9
30 80 40 10
25 80 36 11
20 80 32 12
15 80 28 13
10 80 24 14
5 80 20 15
0 80 16 16

What this means is that until your QB reaches an accuracy of 80, your player's position experience will improve your accuracy skill. At 80, there is a net of 0. And after 80, your accuracy will actually be negatively impacted by your player not having full experience.

This formula holds true for whatever the experience ceiling is. Your QBs accuracy is much more negatively impacted the worse your experience is, which makes sense (at least in this sense.) So if the experience ceiling is set to 66, which is probably closer to what a typical QBs experience level is, you reach net 0 for accuracy at 66. A 100 accuracy QB with 66 experience has an adjusted accuracy of 93.2.

Until we have a better sense of how likely a player will obtain greater than 80 experience, how does everyone feel about that number? If we like it, then the lesson here is that QBs with an accuracy over 80 are less accurate then their attribute says.
Last edited at 10/21/2016 8:20 pm

Re: QB Accuracy With Maths Because I like Maths

By setherick
10/21/2016 7:57 pm
This was meant for the 75 league forums. It can be deleted here.

Re: QB Accuracy With Maths Because I like Maths

By raymattison21
10/21/2016 9:53 pm
"By raymattison21
8/07/2015 11:15 am
Hdurosbdheh:
GG NO.


GG for you. Could not finish watching it. Very rough. Next game I maxed out everything. New Orleans is an experiment testing the limits of the engine. Soon I'll add in a deep pass only, cause this team stinks. Its seems junk teams are decided by the long passes and Special teams"



Check the date and deep passes only. I think this is more of a WR/DB thing. You just need an accurate QB also.

Punt returns are better, but that took too long. Its about time.

Re: QB Accuracy With Maths Because I like Maths

By Infinity on Trial
10/23/2016 5:38 pm
There are a couple of things we don't still don't know here that could explain an awful lot.

1. How low is the average experience figure likely to be? I think 80 may be generous in the data set above.

2. Just how much of a drag are each of these parts of the equation: "(reduction based on scramble skill/movement) x (reduction based on duress) x (reduction based on distance)". If it adds up to a 1% reduction, that's one thing. If it's more like 10% apiece, it's no wonder passing is fubar.

From what we (think we) know, this seems like a reasonable equation. We can quibble over the specifics, but I like that there are some variables. As a Chiefs fan, I would point to a player like Matt Cassel — coaches would rave about how accurate he was in practice, but he was wildly erratic when it counted. It didn't have anything to do with scrambling ability or distance, but I digress. Some system is needed, and at least we (kind of) know what it is now.

My instinct says accuracy should be better in this game, but what I'm really thinking about is bad completion percentages, which are affected by a number of things beyond how a QB throws the ball. The bigger problems affecting passing may be where the QBs are looking, whether they decide to throw the ball in the first place, and defenders with 12-foot-long arms. I can't hardly tell anymore.

Re: QB Accuracy With Maths Because I like Maths

By setherick
10/23/2016 5:46 pm
Here's how I arrived at a ceiling of 80. I hypothesized the experience breakdown like this:

5 x Player Card
2 x Fatigue
1 x Offensive Knowledge
1 x Defensive Knowledge
1 x Crowd Noise

Then I inserted some numbers which seemed reasonable.

5 x 100
2 x 75
1 x 100
1 x 50
1 x 0

This adds up to 800/1000 or 80. I don't know if these numbers are anywhere close to accurate, but I think they're probably a close approximation knowing what I do know about how math works in this game.

Re: QB Accuracy With Maths Because I like Maths

By Infinity on Trial
10/23/2016 5:57 pm
There's a lot of room for variance here — I would guess the overall ceiling is more likely to be 75, and that it often ranges far less. Offensive play knowledge is often in the 20-90 range, depending on how often you've run a given play, and defensive play knowledge seems to be more like 10-40. A roll of the dice can land you a -10 for crowd noise.

Re: QB Accuracy With Maths Because I like Maths

By setherick
10/23/2016 6:01 pm
You and I are defining ceiling differently. I'm using it as an absolute max. I would put the working range between 50 and 65. I assume that a QB's positional rating will account for half, and I assume that within a few preseason games, the QB will have a positional rating of 100.

Here are those numbers so you can stop being so pushy and go back to coding your works CMS interface.

Accuracy Experience Adj Accuracy Difference
100 50 90 -10
95 50 86 -9
90 50 82 -8
85 50 78 -7
80 50 74 -6
75 50 70 -5
70 50 66 -4
65 50 62 -3
60 50 58 -2
55 50 54 -1
50 50 50 0
45 50 46 1
40 50 42 2
35 50 38 3
30 50 34 4
25 50 30 5
20 50 26 6
15 50 22 7
10 50 18 8
5 50 14 9
0 50 10 10

Accuracy Experience Adj Accuracy Difference
100 65 93 -7
95 65 89 -6
90 65 85 -5
85 65 81 -4
80 65 77 -3
75 65 73 -2
70 65 69 -1
65 65 65 0
60 65 61 1
55 65 57 2
50 65 53 3
45 65 49 4
40 65 45 5
35 65 41 6
30 65 37 7
25 65 33 8
20 65 29 9
15 65 25 10
10 65 21 11
5 65 17 12
0 65 13 13

The bigger question is how much FOV influences the initial decision. If 100 FOV always finds the right spot, then a QB with 100 FOV and 50 accuracy is as good with a play he doesn't know as a QB with 50 FOV and 100 Accuracy throwing a play that he has perfect knowledge of.
Last edited at 10/23/2016 6:04 pm

Re: QB Accuracy With Maths Because I like Maths

By Infinity on Trial
10/23/2016 6:03 pm
setherick wrote:
You and I are defining ceiling differently. I'm using it as an absolute max. I would put the working range between 50 and 65. I assume that a QB's positional rating will account for half, and I assume that within a few preseason games, the QB will have a positional rating of 100.


I don't think we are. You put the ceiling at 80, which I think is generous based on typical play knowledge.

Re: QB Accuracy With Maths Because I like Maths

By setherick
10/23/2016 6:05 pm
Infinity on Trial wrote:
setherick wrote:
You and I are defining ceiling differently. I'm using it as an absolute max. I would put the working range between 50 and 65. I assume that a QB's positional rating will account for half, and I assume that within a few preseason games, the QB will have a positional rating of 100.


I don't think we are. You put the ceiling at 80, which I think is generous based on typical play knowledge.


Play knowledge here is the individual's play knowledge and not the team's play knowledge. Don't ask me how those two things are related. I...don't...want...to...know...

Re: QB Accuracy With Maths Because I like Maths

By jdavidbakr - Site Admin
10/24/2016 10:50 am
setherick wrote:
Play knowledge here is the individual's play knowledge and not the team's play knowledge. Don't ask me how those two things are related. I...don't...want...to...know...


The team's play knowledge is just the average of your players' knowledge, but in the game each player's individual knowledge is used. The longer a player has been on your team, the closer his individual play knowledge will be to the team average.